A Improving Performance, Perceived Usability, and Aesthetics with Culturally Adaptive User Interfaces1 KATHARINA REINECKE, University of Zurich ABRAHAM BERNSTEIN, University of Zurich When we investigate the usability and aesthetics of user interfaces, we rarely take into account that what users perceive as beautiful and usable strongly depends on their cultural background. In this paper, we ar- gue that it is not feasible to design one interface that appeals to all users of an increasingly global audience. Instead, we propose to design culturally adaptive systems, which automatically generate personalized inter- faces that correspond to cultural preferences. In an evaluation of one such system, we demonstrate that a majority of international participants preferred their personalized versions over a non-adapted interface of the same web site. Results show that users were 22% faster using the culturally adapted interface, needed less clicks, and made fewer errors, in line with subjective results demonstrating that they found the adapted version significantly easier to use. Our findings show that interfaces that adapt to cultural preferences can immensely increase the user experience. Categories and Subject Descriptors: H.5.2 [Information Interfaces and Presentation]: User Interfaces— User-centered design General Terms: Human Factors, Design, Experimentation Additional Key Words and Phrases: Cultural Adaptivity, Localization, MOCCA, Aesthetics, Usability, User Experience 1. INTRODUCTION With the exponentially growing number of web pages on the internet, searching for the best information increasingly becomes a search for the best information presentation. Many sites simply contradict one’s personal understanding of ‘good design’. But even worse, bad design often goes hand in hand with bad usability. Such findings have led to a growing awareness that aesthetic design is a key component of usability [Hassenzahl 2004; Norman 2004], and a decisive aspect for marketplace success [Bloch 1995]. If neglected, many users rightly decide on another, more attractive and usable web site offering similar content [Lindgaard and Dudek 2003]. With this in mind, research has long discussed the magic formula for a ‘perfect design’, and tried to define what is perceived as beautiful and usable. To some extent, it seems that we can indeed generalize what users consider usable and attractive (e.g., when designing according to the laws of Gestalt Psychology). Some aspects, however, are certainly a matter of personal taste [Norman 2004], or strongly influenced by cultural values [Ito and Nakakoji 1996]. The same conventions apply to user interface preferences, which highly vary across cultures: Asian web sites, for in- stance, are commonly bright and colorful, with frequent animations that try to attract the user’s attention. Their high complexity is often perceived as information overload by Westerners, who prefer more structured content [Marcus and Gould 2001]. Asians, in contrast, have been shown to efficiently filter such dense information [Nisbett 2003]. Users also differ in their design preferences and in their perception of usability at the country level [Barber and Badre 1998; Callahan 2005]. The importance of consid- ering such culturally-determined partialities for a certain look & feel of user interfaces has been demonstrated numerous times. For example, interfaces designed for users of a specific country were perceived as much more attractive [Corbitt et al. 2002], and improved the work efficiency of those they were intended for [Badre 2000; Ford and 1This is a preprint of an article accepted for publication in ACM Transactions on Computer-Human Interaction. A:2 K. Reinecke and A. Bernstein Gelderblom 2003]. Thus, with the Internet being an increasingly global market place, it is unreasonable to design one common web site for everyone, and yet expect to attract an international audience. A more desirable solution would be to design many different versions of a web page in order to cater for users’ cultural preferences. Unfortunately, in this case, designers would soon experience the problem of scale due to the complex and intangi- ble nature of cultural background. To bridge this dichotomy between the need for web sites that cater for individual cultural backgrounds, and an inexpensive method to develop them, we have previously proposed an approach called ‘cultural adaptivity’ [Reinecke and Bernstein 2007]. The idea is that web sites automatically compose personalized interfaces based on the in- dividual cultural backgrounds of its users. We expect cultural adaptivity to improve the performance, as well as the user satis- faction of a personalized web site compared to a non-adapted version of the same site. We refer to performance as the objectively measurable work efficiency, and to user satisfaction as an umbrella term for the users’ subjective perception of the interface, composed of their impressions of both usability, and aesthetics. According to ISO 9241- 11 [1997], usability consists of effectiveness (e.g., a user’s ability to successfully find information and accomplish tasks on a web page), efficiency (he or she can do so in an adequate time frame without frustration), and satisfaction (whether the user en- joys using the web site). Note that rather than being three independently measurable terms, these aspects of usability have been found to interact (cf. [Lindgaard and Dudek 2003]). The user’s judgement of his or her satisfaction, for instance, is thought to be influenced by the perceived usability (as denoted by the effectiveness and efficiency), but also by affective aspects, such as aesthetics [Lindgaard and Dudek 2003]. In the following, we use attractiveness, appeal, and beauty as synonyms for these affective aspects. In an evaluation with 41 users of different cultural backgrounds, we demonstrate the benefit of our approach by showing that the majority preferred their culturally personalized interfaces over a non-adapted version. Moreover, our participants were able to work 22 % faster with the culturally adapted interface, showing that our ap- proach improves both the performance as well as the user satisfaction. In the following, we first discuss the related work, before providing some background on our approach to cultural adaptivity in Section 3. Section 3.2 details the implemen- tation of this approach in our culturally adaptive prototype, which was first introduced in [Reinecke and Bernstein 2009]. The experiment conducted with the system is de- scribed in the subsequent section, followed by a discussion of the implications of this approach for international users. 2. RELATED WORK While both industry and research have long realized that adapting web sites to a user’s culture is a key to market success, efforts in this direction have been limited to soft- ware localization. Localized user interfaces usually modify the most obvious elements to suit a target country and/or region, for instance by adapting to different languages and regional characteristics [Bourges-Waldegg and Scrivener 1998]. Until today, local- ization did not consider the adaptation of user interfaces at presentation level, so that in most cases, the interface’s complexity, colors, and workflows remain unchanged for all users [Kersten et al. 2002]. Despite this fact, research has acknowledged that pro- viding one user interface per country is not enough [Reinecke and Bernstein 2007], due to the difficulty of pinning down cultural background. Indeed, culture does not keep within ‘artificial’ boundaries of countries [Hofstede 1997]. Contrary to this, research has shown that a national interpretation of culture, where the term is equalized with Culturally Adaptive User Interfaces A:3 a specific country, does bundle many user preferences (see, e.g., [Baumgartner 2003] for an overview) — a fact that we have exemplified with web sites of Asians and West- erners in the introduction of this article. Consequently, many researchers have tried to extract design rules by analyzing and comparing web sites (e.g., [Yeo 1996; Barber and Badre 1998; Gould et al. 2000; Zhao et al. 2003; Burgmann et al. 2006]), often resulting in localization guidelines for specific countries, or regions. The implementa- tion of these guidelines in the design of user interfaces is, however, only beneficial for users who (a) belong to a group targeted by the company, and (b) have not been in- fluenced by other countries (and national cultures). An Indian living in Germany, for instance, might have adopted parts of German culture, and her preferences towards web pages have possibly changed. Most likely, she will not feel completely comfortable with German user interfaces that often stand out with a plain design, few colors, and minimal complexity. Despite our increasingly global world, our Indian user and other culturally ambiguous people still have to decide between one interface or the other [Reinecke et al. 2010]. Recognizing this problem, researchers have previously proposed to equip comput- ers with some kind of cultural adaptivity (see, e.g., [O’Neill-Brown 1997] with her call for culturally adaptive agents). The idea is to develop systems that can automatically recognize and adapt themselves to the user’s cultural background. To our knowledge, work in a comparable direction has been limited to research conducted by Kamentz [2006] in the area of e-learning, and by Heimgartner¨ [2005] in the area of cultural adaptivity in navigation systems. Kamentz [2006] relied on a questionnaire to classify the